This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution and sharing with colleagues. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elsevier’s archiving and manuscript policies are encouraged to visit: http://www.elsevier.com/copyright Author's personal copy Journal of Archaeological Science 40 (2013) 960e970 Contents lists available at SciVerse ScienceDirect Journal of Archaeological Science journal homepage: http://www.elsevier.com/locate/jas Characterization of lime carbonates in plasters from Teotihuacan, Mexico: preliminary results of cathodoluminescence and carbon isotope analyses Tatsuya Murakami a, *, Gregory Hodgins b, c, Arleyn W. Simon a a Archaeological Research Institute, School of Human Evolution & Social Change, Arizona State University, P.O. Box 872402, Tempe, AZ 85287-2402, USA NSF-Arizona AMS Facility, Department of Physics, University of Arizona, 1118 E Fourth Street, Tucson, AZ 85721, USA c School of Anthropology, University of Arizona, Haury Building, Tucson, AZ 85721, USA b a r t i c l e i n f o a b s t r a c t Article history: Received 24 February 2012 Received in revised form 23 August 2012 Accepted 29 August 2012 This study characterizes the degree of calcination of lime in lime plaster samples from Teotihuacan, the capital of a regional state in prehispanic Central Mexico. Lime plaster production consists of multiple steps, from the firing of raw materials to the mixing of lime and aggregate and the final application. While previous studies have focused on the compositional variability, specifically the recipe of lime plasters and mortars, the characterization of lime itself has not been sufficiently addressed. In this study, cathodoluminescence analysis coupled with petrographic and image analyses were employed to examine the degree of calcination of lime. The results of cathodoluminescence petrography were further examined through stable carbon isotope and 14C measurements. It appeared that the results of cathodoluminescence analysis are consistent with those of other analytical methods and that there are diachronic changes in the degree of calcination of lime among lime plaster samples. This implies changes in the organization of lime production, specifically the consistency in the control of firing temperature. Ó 2012 Elsevier Ltd. All rights reserved. Keywords: Cathodoluminescence petrography Image analysis Carbon isotopes Radiocarbon measurements Lime plaster Firing techniques Teotihuacan Mesoamerica 1. Introduction Lime plaster production consists of several steps, and thus there are different points at which technical choices can become evident in the final product. Lime plaster production begins by firing raw materials (shell or limestone, CaCO3) to ca. 900  C. This reaction drives off carbon dioxide and produces quicklime (CaO). The reaction may not go to completion so a mixture of calcium oxide and incompletely calcined limestone may result. Water added to quicklime forms a hydroxide paste known as slaked lime (Ca(OH)2). The slaked lime is generally mixed with an aggregate, such as sand or crushed limestone, to create the plaster paste. When the paste is applied to buildings and other features and exposed to the air, the slaked lime absorbs atmospheric CO2 and reforms a calcium carbonate matrix. Previous compositional studies of lime plasters and mortars have shown the technological variability resulting from differential * Corresponding author. Current address: Department of Anthropology, University of South Florida, 4202 E. Fowler Ave., SOC107, Tampa, FL 33620-8100, USA. Tel.: þ1 813 974 2138. E-mail address: tmurakami@usf.edu (T. Murakami). 0305-4403/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jas.2012.08.045 choices at these different production steps (e.g., Carò et al., 2008; Casadio et al., 2005; Hansen, 2000; Littman, 1958; Murakami et al., 2006; Spensley, 2004). Specifically, they have focused on the recipe, including the identification of aggregate and other additives and the ratio of lime and aggregate, among other things. However, the variability in the first step of lime production (i.e., firing of raw materials) has not been explored sufficiently in the previous studies. This is mainly due to the difficulty in distinguishing different phases of calcium carbonate. The degree of calcination (the degree of CO2 removal) is closely related to the quality of lime (Spensley, 2004). Good quality lime can be defined as highly calcined lime and bad quality lime as incompletely calcined lime. The degree of calcination is a function of heating temperature and to keep a higher temperature needs more and/or higher quality fuels and skills. Thus, highly calcined lime reflects more investments of material and/or human resources, and the analysis of the degree of calcination provides useful information on the organization of lime production. This study seeks to characterize the calcium carbonate content in lime plasters based on cathodoluminescence (CL) petrography and other techniques. CL petrography is a common tool for investigating carbonate rocks in geology and other related fields (Machel, 2000) and has Author's personal copy T. Murakami et al. / Journal of Archaeological Science 40 (2013) 960e970 also been applied to the characterization of lime plasters and mortars in archeology (Al-Bashaireh, 2008; Hale et al., 2003; Heinemeier et al., 1997; Lindroos, 2005; Lindroos et al., 2007; see Barbin et al., 1992; Picouet et al., 1999 for CL analysis of other materials). These studies show that CL is a useful tool for identifying different carbonate phases in lime plasters and mortars, specifically unburnt limestone, burnt limestone, incompletely burnt limestone, and lime lumps. While CL was originally adopted as a screening method for selecting samples for radiocarbon dating of calcium carbonates in plasters and mortars, CL analysis provides useful information on the compositional variability and organization of lime plaster and mortar production (Al-Bashaireh, 2008; Murakami, 2010). The presence and proportion of different carbonate phases are often difficult or time-consuming to detect in petrographic thin-section analysis and other kinds of analytical techniques, and CL petrography stands as one of the most efficient methods in this respect. In this paper we present preliminary results of CL analysis on lime plaster samples from Teotihuacan and examine the validity of CL analysis based on other analytical techniques, including stable carbon isotope and 14C measurements. Then, we discuss the implication of the results for the organization of lime production. 961 2. Lime plaster production at Teotihuacan 2.1. Teotihuacan history Teotihuacan is located in the northeastern portion of the Basin of Mexico, within a valley surrounded by volcanoes and mountain ranges (Fig. 1). Teotihuacan developed as an urban center during the Patlachique phase (ca. 150 B.C.eA.D. 1) and was established as the capital of a regional state in Central Mexico during the Tzacualli (ca. A.D. 1e150) or Miccaotli phase (ca. A.D. 150e250) (Cowgill, 1997, 2000; Millon, 1981; Murakami, 2010; Smith and Montiel, 2001). The city of Teotihuacan is characterized by its gigantic monumental structures and highly dense settlement covering ca. 20 square km (Millon, 1973). Major state buildings, including the Moon and Sun Pyramids and the Feathered Serpent Pyramid, were built (and rebuilt) along a central street (the Street of the Dead) in the Miccaotli to Early Tlamimilolpa phases (ca. A.D. 250e300). From the third century A.D. onward, most city residents, probably around 100,000 people, resided in ca. 2300 apartment compounds, distinct walled residential compounds consisting of multiple courtyard units (Cowgill, 2000). Most apartment compounds were built and subsequently rebuilt several times during the Late Tlamimilolpa (ca. A.D. 300e350), Xolalpan (ca. A.D. 350e550), and Fig. 1. Map of Central Mexico, showing the location of Teotihuacan and other important sites. Author's personal copy 962 T. Murakami et al. / Journal of Archaeological Science 40 (2013) 960e970 Metepec (ca. A.D. 550e650) phases until the state collapsed around A.D. 650. 2.2. Organization of lime plaster production at Teotihuacan Lime plaster has been reported from nearly all the excavated architectural complexes at Teotihuacan, although some restricted use can be observed within lower status apartment compounds. Lime plaster was used to coat virtually all the architectural elements, including ceilings, walls, floors, and stairs. As mentioned above, lime plaster production consists of several steps: procurement of raw material, firing of limestone to make lime, preparation of aggregate, slaking lime, mixing of slaked lime and aggregate, and application to walls and floors. These various steps do not have to be completed by the same individual. It is likely that procurement and processing of raw materials are done separately from the final plaster preparation and application at Teotihuacan. Lime production begins by heating raw materials containing calcium carbonate (CaCO3) such as limestone to ca. 900  C to drive off carbon dioxide to produce calcium oxide (CaO) or quicklime. Previous assessments of lime plaster production have found that there were not enough fuel resources to burn limestone within the Teotihuacan valley, and it is likely that limestone was burnt in the source areas and processed quicklime was imported from there (Barba and Córdova, 1999). Barba et al.’s (2009) provenance study of lime indicates that the Tula region was one of the lime sources. In addition, the Zumpango region in the northwestern Basin of Mexico might have been another major source area for lime (see Parsons and Gorenflo, 2008). The second step in lime production is to create a mixture of quicklime and water, known as slaked lime or hydrated lime (carbon hydroxide, Ca(OH)2). Organic and inorganic additives may be mixed with the lime paste. Organic additives include plants (e.g., straw), extracts of plants (e.g., sugars, gums), and animal products (e.g., milk, fats) (Boynton, 1980; Hansen, 2000: 67e68). Organic additives have effects on solubility of calcium hydroxide in water. At Teotihuacan, Torres Montes et al. (2005) suggest that nopal juice was mixed in the plaster matrix. Inorganic additives include various types of clays (aluminosilicates). Lime can react with silica contained in clay minerals and form complex calcium silicates under normal climatic temperature and moisture conditions (Boynton, 1980: 221; Hansen, 2000: 69). At Teotihuacan, a trace amount of clay has been identified in some lime plasters (Barba et al., 2009), although it is not clear whether it was intentionally added or accidentally mixed at some point during the preparation of the lime paste. Since lime paste shrinks to a great extent during drying, which causes extensive cracking, a charge of mineral material or aggregate, such as sand and crushed limestone, is a necessary component of lime plaster (Hansen, 2000: 70e71; Swallow and Carrington, 1995). At Teotihuacan, volcanic ash, quartz, feldspar, amphibole, and limestone were mainly used for aggregate and there were temporal changes (Magaloni, 1996; Magaloni et al., 1992; Murakami, 2010). Aggregate may be added to the lime paste during the slaking of lime or just before its application. The proportion of aggregate to lime binder affects the overall strength of lime plaster (Hansen, 2000: 71), although it is not straightforward to equate different ratios of binder and aggregate to the quality of lime plaster. While a certain amount of aggregate is necessary to make a good-quality lime plaster, aggregate may also be used to expand the paste especially when access to lime is limited (Hansen, 2000; Spensley, 2004). At Teotihuacan, a general decrease in the proportion of lime binder has been noted (Magaloni et al., 1992; Murakami et al., 2006; but see Murakami, 2010). Finally, lime plaster was applied in a thin layer (usually a few millimeters) on the surface of structures at Teotihuacan. When the plaster paste is spread to the building surface, it is exposed to air where it absorbs atmospheric CO2 to reform calcium carbonate, thus forming a hardened plaster surface. However, partly depending on the thickness of plasters and mortars, carbonation process may not complete and calcium oxide/hydroxide may remain in the final product (Goodall et al., 2007; Lawrence et al., 2006). Lime plasters at Teotihuacan are very thin measuring from less than 1 mm up to around 5 mm, so it is likely that there are few uncarbonated lime particles. Based on the comparison of lime plasters (n ¼ 123) from a sample of architectural complexes built and rebuilt in different phases, including major pyramids and apartment compounds, Murakami (2010) suggests that lime plaster production was centrally organized by the state or a group closely related to (or under the control of) the state, during the Miccaotli through Early Xolalpan phases. Murakami found that the composition of lime plaster was highly homogeneous throughout the city. During the Miccaotli and Tlamimilolpa phases, lime plasters were nearly devoid of an aggregate other than the occasional inclusion of crushed limestone fragments. In the Early Xolalpan phase, there was a sudden change in the composition; volcanic ash was introduced as the main aggregate. The fact that this change occurred throughout the city attests to the central organization of lime plaster production. The organization of lime plaster production underwent significant changes by the Metepec phase. It is likely that there were multiple producers who selected different raw materials, and there was no central control on the composition of lime plasters. 3. Materials and methods 3.1. Materials Four lime plaster samples and two limestone samples are examined in this study. Lime plaster samples are from different structures built at different phases (Fig. 2). The first two lime plaster samples (MP125 and 6G1) are from the Moon Pyramid Complex at the northern terminus of the Street of the Dead. MP125 was obtained from the Moon Pyramid. There are seven construction phases at the Moon Pyramid, and the sample is from the wall of the seventh and last construction phase, probably during the Early Xolalpan phase (ca. A.D. 350e450) (see Murakami, 2010; cf. Sugiyama, 2004; Sugiyama and Cabrera, 2007). The sample 6G1 was collected at Complex 6: N5W1, an architectural complex on the west side of the Moon Pyramid. The sample is from the wall of the central temple (Str. 6B), which was built during the Early Xolalpan phase (Carballo, 2005). Both samples were collected during the Moon Pyramid Project, directed by Saburo Sugiyama and Rubén Cabrera. Two other plaster samples (LV2 and LV3) are from a residential apartment compound called La Ventilla II. Both samples were collected in the Project La Ventilla 2004 directed by Rubén Cabrera. LV2 is from a floor recovered in stratigraphic pit excavation (Pit 3) in the Red Borders Unit in the southwestern portion of the compound. It is likely that the floor belongs to the Late Xolalpan phase (ca. A.D. 450e550) (Cabrera, 2003; Cabrera and Gómez, 2008; Gómez and Padilla, 1998: 218). LV3 is from a floor in the West Plaza in the central section of the compound. The floor was recovered from stratigraphic pit excavation (Pit 5) and is tentatively dated to the Late Tlamimilolpa to Early Xolalpan phases (Cabrera, 2003; Cabrera and Gómez, 2008; Gómez and Padilla, 1998: 218). Limestone samples were collected from a surface outcrop in the Tula region and Tepeaca (Tula 2 and Tep4), potential sources for lime at Teotihuacan. As mentioned above, the Tula source is the Author's personal copy T. Murakami et al. / Journal of Archaeological Science 40 (2013) 960e970 963 Fig. 2. The central portion of the city of Teotihuacan, showing the sampling locations. best candidate for lime source (Barba et al., 2009), but Tepeaca limestone may have been used in addition to lime from Tula. The sample Tula 2 was collected from a modern quarry in Tula de Atotonilco. Tepeaca sample was collected from Cerro Tepayacatl. Our preliminary characterization of geological samples using X-ray diffraction and proton-induced X-ray emission revealed that limestone from both sources consists of calcite with high calcium content (Table 1; see also Murakami et al., 2006, 2009). All the Tula samples contain quartz and some have dolomite. 3.2. Cathodoluminescence analysis Cathodoluminescence (CL) analysis of minerals is based on visible luminescence (photons) emitted by a mineral when it is bombarded by high-energy electrons that are produced in a cathode (Marshall, 1988). Emitted light includes photons of various wavelengths, which, along with the intensity of light emission, serve to characterize the mineral and the distribution of some impurities within it. The process of light emission is based on changes in the energy level of a valence electron. When irradiated by an electron beam, a valence electron in a crystal is excited to a state of higher energy and then returns to its original state, emitting a photon. The wavelength of the emitted photon depends on the energy level difference between these different states. It is commonly understood that CL is not emitted by pure crystals but is enhanced by some intrinsic defects (e.g., imperfect structures) and impurities (Marshall, 1988). In particular, many different impurities produce characteristic CL. These impurities Author's personal copy 964 T. Murakami et al. / Journal of Archaeological Science 40 (2013) 960e970 Table 1 Preliminary results of XRD and PIXE analyses of geological limestone samples from the Tula region and Tepeaca (chemical data are normalized to 100%). Sample ID Mineralogy SiO2 Al2O3 MgO CaO P2O5 SO2 SrO Total Tula 1 Tula 2 Tula 4 Tula 5 Tula 7 Tula 8 Tula 10 Tula 11 Tep 1 Tep 2 Tep 3 Tep 4 Tep 5 Calcite, quartz Dolomite, quartz, calcite Calcite, quartz Calcite, quartz, dolomite Calcite, quartz Calcite, quartz Calcite, quartz Calcite, quartz Calcite, quartz Calcite Calcite Calcite Calcite 3.97 21.01 3.76 16.63 2.11 1.69 1.83 10.66 9.04 1.01 0.00 1.78 0.00 1.00 0.29 0.00 0.00 0.00 0.00 0.30 0.62 0.61 0.32 0.00 0.00 0.00 0.68 26.39 0.72 2.73 0.73 0.48 0.51 0.82 0.41 0.39 0.40 0.32 0.67 94.00 52.12 95.31 80.47 96.92 97.60 96.90 87.63 89.71 97.86 98.85 97.78 98.70 0.00 0.00 0.00 0.00 0.00 0.00 0.25 0.00 0.00 0.23 0.60 0.00 0.42 0.31 0.17 0.16 0.17 0.23 0.19 0.18 0.23 0.20 0.16 0.15 0.12 0.18 0.04 0.02 0.05 0.00 0.00 0.04 0.03 0.04 0.03 0.04 0.00 0.00 0.04 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 are called “activators” (Marshall, 1988: 9). The most common activators are the transition elements, rare earth elements (REEs), and actinides. It should be noted that a few elements can inhibit or weaken the CL emission. These elements are called quenchers (Marshall, 1988: 11). The most common quencher is Fe2þ. The presence of different activators and quenchers, along with their position in a crystal structure, determines the CL of a mineral. In CL of lime plaster and mortar, Mn2þ is the only activator coupled with Fe2þ quenching and the CL intensity is basically a function of the ratio Mn2þ/Fe2þ in solid solution (Habermann et al., 2000; Lindroos, 2005). In addition, changing pH conditions during plaster and mortar hardening are a factor for a variety of CL (Machel, 2000). Previous studies (Lindroos, 2005; Lindroos et al., 2007; AlBashaireh, 2008) demonstrate that some unburnt limestone has a strong red-orange luminescence, whereas lime binder or burnt limestone produced from a luminescent parent shows a dull tilered to brown luminescence. There are both luminescent and dull lime lumps, depending on the ratio of Mn2þ/Fe2þ (Lindroos, 2005: 50). Incompletely burnt limestone often shows dark red uneven luminescence (Lindroos, 2005: 51). In addition to calcium carbonate, other minerals in plaster thin sections also show luminescence: quartz has a blue luminescence while feldspars (especially K-feldspars) have a green luminescence (Lindroos, 2005). However, these two classes of minerals may not be clearly distinguished because quartz may show other color ranges, including green (Richter et al., 2003: 128) and some types of feldspar have blue emissions (Richter et al., 2003: 137; see also Götze, 2000). CL analysis was conducted on thin-sections of lime plaster samples. Thin-sections were made at Spectrum Petrographics, Inc., where the samples were first vacuum-impregnated with epoxy (EPOTEK 301 mixed with quartz sand). They were cut and mounted on a glass slide and then were ground to 30 mm thickness. We used CL microscopy (Relion Industries Cathodoluminescence System) attached to a regular reflected light microscope at the NSF-Arizona AMS Facility, University of Arizona. Thin-sections are put in a low vacuum chamber set on the stage of a microscope and are irradiated by accelerated electrons (20e25 keV and under 1 mA) generated by a cathode gun attached to the vacuum chamber. The resultant CL image is captured by a digital camera (Olympus DP71) attached to the microscope. In addition, thin-sections were observed under a petrographic (polarizing) microscope to evaluate the results of CL analysis. CL micrographs were inspected to identify the composition of the plaster matrix (binder and aggregate). We conducted an image analysis on CL micrographs to quantify the degree of calcination and the proportion of each component (e.g., quartz and feldspar sands) (see Lindroos, 2005). A free image analysis program called ImageJ 1.42q (Wayne Rasband, National Institutes of Health; http:// rsbweb.nih.gov/ij/index.html) was used. In this program, the RBG colors were separated and the proportion of each component was calculated (blue and green were merged because both quartz and feldspar may show these color ranges; Fig. 3). In addition, micrographs taken under a petrographic microscope with plain polarized light were analyzed to quantify the proportion of lime binder. Most aggregate particles are semi-translucent to completely translucent in our samples, and images were converted to black and white and the black area was measured (Fig. 4). 3.3. Analysis of carbon isotopes We measured the delta (d) 13C values (ratio of sample 13Ce12C compared to the 13Ce12C of a standard) on all the samples and 14C content on a subset of samples. Pachiaude et al. (1986) investigated carbon isotope fractionation during CO2 absorption by CaO. Under laboratory conditions, the newly formed carbonate had a d13C value of 21&. Van Strydonck et al. (1986) suggested that the d13C value of CO2 liberated during mortar hydrolysis would indicate whether the source carbonate was originally derived from the atmosphere or limestone. However, in further experiments (Van Strydonck et al., 1989) demonstrated that 13C depletion in mortars and plasters was a consequence of non-equilibrium conditions during the carbonization reaction, and so cannot provide an indication of fossil source. Nevertheless, marine carbonate and limestone have the d13C values close to 0&, and pure lime plasters and mortars have d13C values much lower than this. Previously Murakami et al. (2009) found that the d13C value of incompletely calcined limestone is dependent on the degree of calcination. As for the 14C content, limestone is devoid of 14C due to its geological age (Tertiary to Cretaceous in Central Mexico), while newly crystallized calcium carbonate should contain 14C at the time of its application since calcium oxide absorbs atmospheric CO2 to reform calcium carbonate crystals. Thus, older dates than expected indicate that lime plaster contains 14C-dead CO2 from limestone and/or incompletely calcined limestone. The measurement of both d13C and 14C values was originally conducted for radiocarbon dating of lime plasters and thus sample preparation followed the procedures developed for that purpose (Hodgins et al., 2006; Sonninen and Jungner, 2001; Heinemeier et al., 1997). The samples were first surface-cleaned using a razor blade or Dremel tool. Plaster samples were gently crushed using a mortar and pestle. Limestone required vigorous crushing and milling. Crushed samples were transferred to a glove box and wetsieved under a nitrogen atmosphere, using helium degassed water. The 45e63 micron-sized fractions were isolated and dried at 65  C under nitrogen in preparation for acid hydrolysis. One hundred milligram aliquots of powdered sample were subjected to acid hydrolysis under vacuum as a time-course reaction, initiated by the Author's personal copy T. Murakami et al. / Journal of Archaeological Science 40 (2013) 960e970 965 Fig. 3. Procedure of the image analysis of a CL micrograph (sample LV2). The original image (top-left) is split into three color components: red (R), green (G), and blue (B). Each image is then converted to black and white and its proportion is calculated. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) addition of 4 ml of degassed 85% phosphoric acid. Evolved CO2 was recovered into a sequence of ampoules at 0.5, 1, 2, 3, 5, 10, 20, 40, 80, and 160 min. The stable isotopic composition of the evolved gases was measured on a VG Isotech stable isotope mass spectrometer. Radiocarbon measurements were carried out by accelerator mass spectrometry using established graphitization and measurement methods. 4. Results 4.1. CL petrography CL analysis shows that limestone samples from both the Tula region and Tepeaca have a bright red luminescence, which makes CL analysis of plasters derived from them possible, and confirms the previous studies (Fig. 5). CL micrographs of archaeological samples show the presence of different carbonate phases along with aggregate particles (Fig. 6). Lime binder and lime lumps have a dull-purple luminescence. They can be distinguished by their tone of color; lime binder has darker tone. A strong red-orange (sometimes with yellow) luminescence indicates incompletely calcined and/or unburnt limestone fragments. Observations under a polarizing microscope revealed that these particles are not unburnt limestone as suggested by Magaloni (1996); Magaloni et al. (1992), but incompletely calcined limestone fragments (Fig. 7). Spensley (2004) demonstrates that gray color is characteristic of incompletely calcined lime under plain-polarized light, while brownish yellow is characteristic of calcined lime. The strongly luminescent particles are gray under plain-polarized light, which supports the interpretation that those particles are incompletely calcined limestone fragments. Other than carbonate phases, we also identified quartz/feldspar with blue and green luminescence and amorphous particles with no luminescence, which were identified as volcanic ash through petrographic analysis. Epoxy shows no luminescence and quartz sands mixed in epoxy are light blue and green. Fig. 4. Image analysis of a plain polarized micrograph (LV2). The original image (left) is converted to a black and white image (right), and the black area (binder) is quantified. Author's personal copy 966 T. Murakami et al. / Journal of Archaeological Science 40 (2013) 960e970 Fig. 5. CL micrographs of limestone samples from the Tula region (left) and Tepeaca (right). Fig. 6. Plain polarized (left) and CL (right) micrographs of the analyzed samples. Author's personal copy 967 T. Murakami et al. / Journal of Archaeological Science 40 (2013) 960e970 Fig. 7. Plain polarized micrographs of incompletely calcined limestone fragments (left: 6G1; right: LV2). There is compositional variability among archaeological samples (Table 1), and the proportion of different carbonate phases considerably varies. The sample MP125 contains very few incompletely calcined limestone fragments (0.48% of the total plaster matrix or 0.55% of the binder), whereas 6G1 and LV2 have a larger amount of incompletely calcined limestone fragments; 1.99% (or 2.41% of the binder) for 6G1 and 3.0% (3.56% of the binder) for LV2 (Fig. 6). The sample 6G1 has a differential distribution of incompletely calcined limestone, and the near surface area has 1.01% (or 1.2% of the binder) and the lower portion has 2.73% (or 3.34% of the binder). The amount of incompletely calcined limestone fragments for LV3 (Fig. 6) is somewhere between those for MP125 and 6G1/LV2 (1.23% of plaster or 1.44% of the binder). Besides the carbonate phases, there is also variation in the proportion of different kinds of aggregate (Table 2). 4.2. Stable carbon isotope and 14 C measurements The stable carbon isotope analysis shows the variation in the d13C values among samples (Fig. 8) and supports the interpretation of CL analysis. As expected, limestone samples have the d13C values close to 0& and lime plaster samples have lower values. MP125 has the lowest d13C values (around 14& to 15&), and only a slight increase was observed in the time course reaction. This indicates that MP125 contains very few of incompletely calcined limestone and/or unburnt limestone at least in the micron-level size fraction. Since some incompletely calcined limestone fragments were identified in CL analysis, it is likely that these fragments were larger than 63 mm and thus excluded by the sieving steps in the stable isotope and radiocarbon sample preparation protocols. In contrast to MP125, the samples 6G1 and LV2 have higher d13C values at around 11&, which increase after 40 min to reach around 5&. Table 2 Proportion of different constituents of lime plaster samples. Sample Spot Lime binder Aggregate Calcined (dull red)a Incompletely calcined (red) Binder total Limestone (red)b Quartz/feldspar (blue/green) Volcanic ash etc. (black)c Aggregate total MP125 1 2 3 Ave. 89.65 80.68 85.58 85.30 1.03 0.13 0.27 0.48 90.68 80.81 85.84 85.78 0.00 3.81 0.00 1.27 0.29 0.17 1.06 0.51 9.03 15.21 13.10 12.45 9.32 19.19 14.16 14.22 6G1 (near surface) 1 2 3 Ave. 82.40 83.65 84.08 83.38 1.20 1.14 0.70 1.01 83.60 84.78 84.78 84.39 0.00 0.00 0.00 0.00 1.17 0.75 0.11 0.67 15.23 14.47 15.11 14.94 16.40 15.22 15.22 15.61 6G1 (lower portion) 4 5 6 7 Ave. 75.10 79.39 78.75 82.74 78.99 3.44 3.04 2.48 1.97 2.73 78.54 82.43 81.23 84.71 81.73 0.00 0.00 0.00 0.00 0.00 1.16 0.34 0.89 0.63 0.76 20.30 17.23 17.88 14.66 17.52 21.46 17.57 18.77 15.29 18.27 6G1 (total) Ave. 80.88 1.99 82.87 0.00 0.71 16.42 17.13 LV2 1 2 3 4 5 Ave. 81.90 83.17 79.38 81.01 80.74 81.24 2.07 3.82 3.09 2.64 3.35 3.00 83.97 86.99 82.47 83.66 84.09 84.23 0.00 0.00 0.00 0.00 0.00 0.00 1.79 0.83 1.69 1.18 1.70 1.44 14.25 12.18 15.85 15.16 14.21 14.33 16.03 13.01 17.53 16.34 15.91 15.77 LV3 1 2 3 Ave. 84.88 84.26 83.98 84.37 0.96 1.70 1.03 1.23 85.84 85.97 85.01 85.60 0.00 0.00 0.00 0.00 2.62 1.99 2.92 2.51 11.55 12.05 12.08 11.89 14.16 14.04 14.99 14.40 a b c The proportion of calcined lime binder was calculated as the total percent of binder subtracted by the percent of red luminescent area. Limestone aggregate was observed only in the sample MP125. Its proportion was calculated independently by the image analysis program ImageJ. The proportion of volcanic ash etc. was calculated as the total percent of aggregate subtracted by the percent of other measured aggregate. Author's personal copy 968 T. Murakami et al. / Journal of Archaeological Science 40 (2013) 960e970 2 5000 0 20 40 60 80 100 120 140 delta C (per mil) -4 Tep4 -6 6G1 LV2 -8 LV3 -10 LV2 4500 160 MP125 -12 -14 Radiocarbon date (y BP) 0 -2 6G1 4000 MP125 3500 3000 2500 2000 1500 -16 0 Time (minutes) Fig. 8. The d13C measurements of evolved CO2 from lime plaster and limestone samples. This is a clear indication that these samples contain incompletely calcined or unburnt limestone fragments, which accords well with the results of CL analysis. The sample LV3 has the d13C values at around 13&, which is between MP125 and 6G1/LV2. It is likely that LV3 has more incompletely calcined or unburnt limestone fragments than MP125, but unlike 6G1 and LV2, there is no significant increase in the d13C values (although the d13C could not be measured for the fraction at 160 min due to the insufficient amount of CO2). During the experiment, it was also noted that the d13C values never reached 0& unlike plaster samples from the Maya region (Hodgins et al., 2006). This indicates that lime plasters from Teotihuacan are devoid of unburnt limestone and further supports the interpretation based on CL and petrographic analyses. It should be noted that initial spikes were observed for all the samples including limestones. It is likely that the 1& depression of the first fraction is an artifact of kinetic fractionation during the initial stages of collection (Murakami et al., 2009). The results of 14C measurements complement those of CL and stable isotope analyses. The radiocarbon date of the sample MP125, measured from the CO2 collected during the first 30 s of hydrolysis, is close to the expected date of construction (Table 3). Two other dates obtained from the same sample using different method of CO2 extraction (Murakami et al., 2009) are even closer to the expected date. As with the stable isotope analysis, it is likely that MP125 contains very few incompletely calcined limestones. However, unlike the stable isotope analysis, 14C measurements clearly indicate that MP125 contains a small amount of incompletely calcined or unburnt limestone fragments. The dates obtained from evolved CO2 after 30 s of hydrolysis are significantly older (around 2600e2700 BP) than expected, which reflects limestone contamination (Fig. 9). The radiocarbon dates obtained from the first 30 s-fraction of 6G1 and LV2 are both 100e200 years older than expected dates (Table 3). The dates from later fractions (after 30 s) have significantly older ages. Thus, as with the stable isotope analysis, 14C measurements clearly indicate that 6G1 and LV2 contain a larger Table 3 Radiocarbon dates of lime plasters and expected dates for each sample. 14 C date MP125 1598  43 BP 6G1 1912  35 BP LV2 1887  36 BP Calibrated (1 sigma) Calibrated (2 sigma) Expected date AD 419e467, AD 481e534 AD 56e129 AD AD AD AD AD AD 350e450 AD 67e140, AD 155e168, AD 195e209 355e366, 381e565 5e178, 190e213 53e227 AD 350e450 AD 450e650 10 20 30 40 50 60 70 80 90 100 Percent hydrolysis Fig. 9. Radiocarbon dates of lime plasters versus percent hydrolysis. amount of incompletely calcined limestone fragments than that for MP125. 5. Discussion 5.1. Validity of CL analysis Stable carbon isotope and 14C measurements support our interpretation based on CL petrography. Although limestone used at Teotihuacan is different in source area, age, and chemical composition than that analyzed by other researchers, our results are consistent with those of Lindroos (2005). However, there is a certain limitation in CL analysis as well. Incompletely calcined limestone fragments had a range of spectra from a strong redorange to dark red or purple luminescence, which are difficult to distinguish from unburnt limestone. Although Lindroos (2005: 51) suggested that uneven luminescence is indicative of incompletely calcined limestone, it appeared that not all the incompletely calcined limestone fragments have an uneven luminescence. In this respect, the results of CL analysis need to be examined under a polarizing microscope. Color distinction (gray versus yellowish brown) under a plain polarized light seems reliable to distinguish incompletely burnt and unburnt limestone (Spensley, 2004). This by no means signifies that CL analysis is ineffective in characterizing lime carbonate. We emphasize that there are certain identifiable differences between incompletely burnt and unburnt limestone, as suggested by Lindroos (2005). It should be understood that the difference between burnt and unburnt limestone is a continuum, which makes it likely that the spectrum of luminescence is also continuous between burnt and unburnt limestone. Further experiments along with microtextural analysis will be necessary to determine the correlation between firing temperature and luminescence patterns. While thermal decomposition of calcite is homogeneous (Rodriguez-Navarro et al., 2009), our preliminary analysis shows some mineralogical and chemical variations among parent limestone samples. Since the firing regimen of limestone is directly related to its chemical composition (e.g., Rodriguez-Navarro et al., 2009), different microtextures might result from different raw materials, which may or may not affect luminescence patterns. Moreover, future research needs to consider other factors, such as the decay of newly formed calcite and the formation of secondary calcite, as possible sources of variations in luminescence patterns. The quantification of incompletely calcined limestone through image analysis turns out to be very useful for characterizing lime plasters from Teotihuacan. This is because petrographic analysis and stable isotope and 14C measurements all suggest that Author's personal copy T. Murakami et al. / Journal of Archaeological Science 40 (2013) 960e970 Teotihuacan lime plasters are almost devoid of unburnt limestone, and particles with a red-orange luminescence could be assigned to incompletely calcined limestone (although other factors should be considered as mentioned above). It should be clear that, when incompletely burnt and unburnt limestone fragments are mixed, it is impossible to isolate one or the other based solely on color spectrum. In this case, the proportion of incompletely burnt and unburnt limestone should be quantified through point-counting methods under a petrographic microscope (Spensley, 2004) or through other methods (Casadio et al., 2005). In sum, while CL analysis coupled with petrographic and image analyses turned out to be the most efficient method for characterizing lime carbonate in Teotihuacan plasters in terms of time and cost, its utility and scope depend on the composition of lime plasters and mortars and need to be judged case by case. 5.2. Firing techniques and the organization of lime production The distinction between incompletely burnt and unburnt limestone has not been a central issue for selecting samples that can be radiocarbon dated and has not been paid much attention since both affect 14C measurements in a similar way. However, this distinction has a great implication for the organization of lime production. We argue that the differential proportion of incompletely calcined limestone is likely related to firing techniques of lime producers, rather than the technical choices of aggregate by lime plasterers. It was observed that micron-sized particles of incompletely calcined limestone are present in all the samples and are well distributed in the plaster matrix especially in LV2 and 6G1. It is likely that they are residues of the same burnt limestone, from which fully calcined lime was derived. This implies that these incompletely burnt limestone fragments were mixed unintentionally, although they (especially large ones) are functionally aggregate. Thus, changes in the proportion of incompletely calcined limestone are closely related to the variability in the skills of lime producers. As mentioned above, the degree of calcination is a function of firing temperature, and lime with few incompletely calcined limestone fragments implies the better control of firing temperature. In this respect, Hansen (2000: 211) provides two alternative methods to remove unburnt material. One is the removal by hand of recognizably unburnt material, and the second is sieving of limeputty. However, given the small size of incompletely burnt limestone fragments, it is unlikely that they can be removed by hand. As for the second possibility, as Hansen mentions, there is no evidence for sieving practice. Therefore, it seems reasonable to suggest that the changing proportion of incompletely canlcined limestone was related to differential firing techniques, although methods proposed by Hansen might have been used in combination. In any case, it is certain that the proportion of incompletely calcined limestone is closely related to skills of lime producers. It is possible that the degree of calcination, and thus the skills of lime producers, changed through time at Teotihuacan. The results of this study indicate that MP125 and LV3 are composed of highly calcined lime, both of which pertain to the Early Xolalpan phase (ca. A.D. 350e450). The sample 6G1 also belongs to a structure built during the Early Xolalpan phase, but this sample is from the outermost layer of several replastered layers and is likely dated to the latter part of the Early Xolalpan or later. LV2 is from a Late Xolalpan to Metepec (ca. A.D. 450e650) context. Thus, the degree of calcination decreased through time at both the Moon Pyramid Complex and the La Ventilla II apartment compound, and it is possible that this change occurred throughout the city. In a separate study, Murakami (2010) conducted petrographic and CL analyses on 123 lime plaster samples from 16 architectural complexes built 969 and rebuilt in different phases and detected that the composition is highly homogeneous among different structures during the Early Xolalpan phase and most samples have very few incompletely calcined limestone fragments. In contrast, samples from the Late Xolalpan and Metepec phases showed variability in both recipe and quality, and some contained a large amount of incompletely calcined limestone fragments. This observation is consistent with changes in the scale of lime production. During the Tlamimilolpa and Early Xolalpan phases, around 2300 apartment compounds were built and rebuilt throughout the city and there was a high demand for lime. The lime production was certainly intensified, which may have led to the specialization of lime producers. The Tula region was under Teotihuacan’s control in these phases, and it is possible that lime producers were centrally controlled through a secondary center (Chingú) of the Teotihuacan state and perhaps other centers (Díaz, 1980; Mastache et al., 2002). During the Late Xolalpan phase, construction activities almost ceased within the civic-ceremonial core of the city and only some apartment compounds were rebuilt (Millon, 1988). The scale of lime production and the number of specialized lime producers should have been significantly reduced, which would have resulted in the production of lower quality lime. Mastache et al. (2002: 59e 60) suggests the weakened state control of the Tula region from the Late Xolalpan phase, which implies that the Teotihuacan state could not longer sustain specialist lime producers. The causes of reduced construction activities and lime production are not wellunderstood, but there is evidence to suggest some political conflict within the city and/or between the Teotihuacan state and its hinterlands (Hirth, 1978; Millon, 1988). 6. Conclusion This preliminary study shows that the results of CL analysis are consistent with those of other analytical techniques. With the aid of a petrographic microscope, CL analysis can distinguish different carbonate phases, specifically calcined and incompletely calcined lime. Moreover, image analysis of CL micrographs is an efficient way to quantify the proportion of incompletely calcined limestone fragments and other materials, as compared to point-counting methods, though with some limitations. Future research should address other sources of variability in luminescence patterns to better understand technological practices of ancient cultures. Overall, the results of this study suggest diachronic changes in the degree of calcination in lime plasters at the Moon Pyramid Complex and the La Ventilla II apartment compound. It is likely that these changes are related to changing organization of lime production in source areas. During the Late Tlamimilolpa and Early Xolalpan phases (ca. A.D. 300e450), intensified demand for lime by city residents resulted in the specialized production of goodquality, highly calcined lime. From the Late Xolalpan phase (ca. A.D. 450e550) onward, due to reduced construction activities within the city, the scale of lime production was significantly reduced and the control of firing temperature became inconsistent resulting in overall decreased degree of calcination of lime. Acknowledgments This study was funded by the National Science Foundation (SBE0514396). We thank Saburo Sugiyama and Rubén Cabrera for their permission to collect lime plaster samples during the Moon Pyramid Project and the Project La Ventilla 2004. We are also grateful to the National Institute of Anthropology and History, Mexico, for their permission to export the samples. Our thanks also go to Alf Lindroos for his guidance for CL analysis. Comments from Author's personal copy 970 T. Murakami et al. / Journal of Archaeological Science 40 (2013) 960e970 two anonymous reviewers greatly helped to refine the scope of the paper and we acknowledge their time and effort. References Al-Bashaireh, K.S., 2008. Chronology and Technological Production Styles of Nabatean and Roman Plasters and Mortars at Petra (Jordan). Ph.D. dissertation, University of Arizona, Tucson. Barba, L., Blancas, J., Manzanilla, L.R., Ortiz, A., Barca, D., Crisci, G.M., Miriello, D., Pecci, A., 2009. Provenance of the limestone used in Teotihuacan (Mexico): a methodological approach. Archaeometry 51, 525e545. Barba P., L.A., Córdova F., J.L., 1999. Estudios energéticos de la producción de cal en tiempos teotihuacanos y sus implicaciones. Latin American Antiquity 10, 168e179. Barbin, V., Ramseyer, K., Decrovez, D., Burns, S.J., Chambay, J., Maier, J.L., 1992. Cathodoluminescence of white marble: an overview. Archaeometry 34, 175e183. Boynton, R.S., 1980. Chemistry and Technology of Lime and Limestone. Wiley and Sons, New York. Cabrera C., R., 2003. El Proyecto Arqueológico La Ventilla 1992-1994. Resumen de sus resultados. In: Serrano S., C. (Ed.), Contextos arqueológicos y osteología del barrio de La Ventilla, Teotihuacan (1992-1994). Instituto de Investigaciones Antropológicas, Universidad Nacional Autónoma de México, Mexico D.F, pp. 19e30. Cabrera C., R., Gómez C., S., 2008. La Ventilla: a model for a barrio in the urban structure of Teotihuacan. In: Mastache, A.G., Cobean, R.H., García Cook, A., Hirth, K.G. (Eds.), Urbanism in Mesoamerica, vol. 2. INAH and Pennsylvania State University, Mexico, D.F. and University Park, pp. 37e83. Carballo, David M., 2005. State Political Authority and Obsidian Craft Production at the Moon Pyramid, Teotihuacan, Mexico. Ph.D. dissertation, University of California, Los Angeles. Carò, F., Riccardi, M.P., Mazzilli Savini, M.T., 2008. Characterization of plasters and mortars as a tool in archaeological studies: the case of Lardirago Castle in Pavia, northern Italy. Archaeometry 50, 85e100. Casadio, F., Chiari, G., Simon, S., 2005. Evaluation of binder/aggregate ratios in archaeological lime mortars with carbonate aggregate: a comparative assessment of chemical, mechanical and microscopic approaches. Archaeometry 47, 671e689. Cowgill, G.L., 1997. State and society at Teotihuacan, Mexico. Annual Review of Anthropology 26, 129e161. Cowgill, G.L., 2000. The Central Mexican Highlands from the rise of Teotihuacan to the decline of Tula. In: Adams, R.E.W., MacLeod, M.J. (Eds.), The Cambridge History of the Native Peoples of the Americas. Mesoamerica, Part 1, vol. II. Cambridge University Press, Cambridge, UK, pp. 250e317. Díaz O., C.L., 1980. Chingú: Un sitio Clásico del área de Tula, Hgo. Colección Científica, núm. 90. Instituto Nacional de Antropología e Historia, Mexico, D.F. Götze, J., 2000. Cathodoluminescence Microscopy and Spectroscopy in Applied Mineralogy. Freiberger Forschungshefte. C485. Gómez C., S., Padilla R., R., 1998. Correlación cronológica de la pintura mural en tres conjuntos arquitectónicos de La Ventilla, Teotihuacán. In: Brambila, R., Cabrera C., R. (Eds.), Los ritmos de cambio en Teotihuacán: reflexiones y discusiones de su cronología. Instituto Nacional de Antropología e Historia, Mexico, D.F., pp. 201e221. Goodall, R.A., Hall, J., Edwards, H.G.M., Sharer, R.J., Viel, R., Fredericks, P.M., 2007. Raman microprobe analysis of stucco samples from the buildings of Maya Classic Copan. Journal of Archaeological Science 34, 666e673. Hale, J., Heinemeier, J., Lancaster, L., Lindroos, A., Ringbom, A., 2003. Dating ancient mortar. American Scientist 91, 130e138. Habermann, D., Neuser, R.D., Richter, K., 2000. Quantitative high resolution spectral analysis of Mn2þ in sedimentary calcite. In: Pagel, M., Barbin, V., Blanc, P., Ohnenstetter, D. (Eds.), Cathodoluminescence in Geosciences. Springer-Verlag, Berlin, pp. 331e358. Hansen, E., 2000. Ancient Maya Burnt-lime Technology: Cultural Implications of Technological Styles. Ph.D. dissertation, University of California, Berkeley. Heinemeier, J., Jungner, H., Lindroos, A., Ringbom, Å., von Konow, T., Rud, N., 1997. AMS 14C dating of lime mortar. Nuclear Instruments and Methods in Physics Research B 123, 487e495. Hirth, K.G., 1978. Teotihuacan regional population administration in eastern Morelos. World Archaeology 9, 320e333. Hodgins, G., Vonarx, A.J., Bachand, B., 2006. Radiocarbon dating of protoclassic lime plasters. In: Pérez-Arantegui, J. (Ed.), Proceedings of the 34th International Symposium on Archaeometry, Zaragoza, 3e7 May 2004. Institución Fernando el Católico, Zaragoza, pp. 73e78. Lawrence, R.M.H., Mays, T.J., Walker, P., Ayala, D.D., 2006. Determination of carbonation profiles in non-hydraulic lime mortars using thermogravimetric analysis. Thermochimica Acta 444, 179e189. Lindroos, A., 2005. Carbonate Phases in Historical Lime Mortars and Pozzolana Concrete: Implications for 14C Dating. Dissertation, Department of Geology and Mineralogy, Åbo Akademi University, Åbo, Finland. Lindroos, A., Heinemeier, J., Ringbom, Å., Braskén, M., Sveinbjörnsdóttir, A., 2007. Mortar dating using AMS 14C and sequential dissolution: examples from Medieval, non-hydraulic lime mortars from the Åland Islands, SW Finland. Radiocarbon 49, 47e67. Littman, E.R., 1958. Ancient Mesoamerican mortars, plasters, and stuccos: the composition and origin of sascab. American Antiquity 24, 172e176. Machel, H.G., 2000. Application of cathodoluminescence to carbonate diagenesis. In: Pagel, M., Barbin, V., Blanc, P., Ohnenstetter, D. (Eds.), Cathodoluminescence in Geosciences. Springer-Verlag, Berlin, pp. 271e301. Magaloni, D., 1996. El espacio pictórico teotihuacano. Tradición y técnica. In: de la Fuente, B. (Ed.), La pintura mural prehispánica en México I: Teotihuacán, Tomo II. Estudios. Universidad Nacional Autónoma de México and Instituto de Investigaciones Estéticas, Mexico, D.F., pp. 187e225. Magaloni, D., Falcon, T., Cama, J., Siegel, R.W., Lee, R., Pancella, R., Baños, L., Castaño, V., 1992. Electron microscopy studies of the chronological sequences of Teotihuacan plaster technique. Material Issues in Art and Archaeology 3, 997e1005. Marshall, D.J., 1988. Cathodoluminescence of Geological Materials. Unwin Hyman, Boston, MA. Mastache, A.G., Cobean, R.H., Healan, D.M., 2002. Ancient Tollan: Tula and the Toltec Heartland. University Press of Colorado, Boulder, CO. Millon, R., 1973. The Teotihuacan Map. In: Part 1: Text. Urbanization at Teotihuacan, Mexico, vol. 1. University of Texas Press, Austin. Millon, R., 1981. Teotihuacan: city, state, and civilization. In: Bricker, V., Sabloff, J.A. (Eds.), Supplement to the Handbook of Middle American Indians. Archaeology, vol. 1. University of Texas Press, Austin, pp. 198e243. Millon, R., 1988. The last years of Teotihuacan dominance. In: Yoffee, N., Cowgill, G.L. (Eds.), The Collapse of Ancient States and Civilizations. University of Arizona Press, Tucson, pp. 102e164. Murakami, T., 2010. Power Relations and Urban Landscape Formation: a Study of Construction Labor and Resources at Teotihuacan. Ph.D. dissertation, Arizona State University, Tempe. Murakami, T., Garraty, C., Simon, A.W., McKelvy, M., 2006. Thermogravimetric analysis (TGA) of archaeological materials, part I: lime plaster. SAS Bulletin 29 (4), 14e17. Murakami, T., Hodgins, G., Vonarx, A.J., Simon, A.W., 2009. Radiocarbon dating of lime carbonates in Mesoamerican plasters: studies from Teotihuacan. In: Moreau, J.-F., Auger, R., Chabot, J., Herzog, A. (Eds.), Proceedings of the 36th International Symposium on Archaeometry, Quebec City, Canada, pp. 83e92. Pachiaude, C., Marechal, J., Van Strydonck, M., Dupas, M., Dauchot-Dehon, M., 1986. Isotopic fractionation of carbon during CO2 absorption by mortar. Radiocarbon 28, 691e697. Parsons, J.R., Gorenflo, L.J., 2008. The patterning of settlement. In: Parsons, J.R. (Ed.), Prehispanic Settlement Patterns in the Northwestern Valley of Mexico: The Zumpango Region. Memoirs, No. 45, Museum of Anthropology. University of Michigan, Ann Arbor, pp. 61e99. Picouet, P., Maggetti, M., Piponnier, D., Schvoerer, M., 1999. Cathodoluminescence spectroscopy of quartz grains as a tool for ceramic provenance. Journal of Archaeological Science 26, 943e949. Richter, D.K., Götte, Th., Götze, J., Neuser, R.D., 2003. Progress in application of cathodoluminescence (CL) in sedimentary petrology. Mineralogy and Petrology 79, 127e166. Rodriguez-Navarro, C., Ruiz-Agudo, E., Luque, A., Rodriguez-Navarro, A.B., OrtegaHuertas, M., 2009. Thermal decomposition of calcite: mechanisms of formation and textural evolution of CaO nanocrystals. American Mineralogist 94, 578e593. Smith, M.E., Montiel, L., 2001. The archaeological study of empires and imperialism in pre-Hispanic Central Mexico. Journal of Anthropological Archaeology 20, 245e284. Sonninen, E., Jungner, H., 2001. An improvement in the preparation of mortars for radiocarbon dating. Radiocarbon 43, 271e273. Spensley, E., 2004. Micromorphology of Construction and Culture at Trinidad de Nosotros, Peten, Guatemala. M.A. thesis, Boston University, Boston, MA. Sugiyama, S., 2004. Governance and policy at Classic Teotihuacan. In: Hendon, J.A., Joyce, R.A. (Eds.), Mesoamerican Archaeology: Theory and Practice. Blackwell Publishing, Oxford, UK, pp. 97e123. Sugiyama, S., Cabrera C., R., 2007. The Moon Pyramid Project and the Teotihuacan polity: a brief summary of the 1998e2004 excavations. Ancient Mesoamerica 18, 109e125. Swallow, P., Carrington, D., 1995. Limes and lime mortars: part one. Journal of Architectural Conservation 1, 7e25. Torres Montes, L., García, M.R., Gazzola, J., Gómez, S., 2005. Analysis of stucco floors from the Citadel of the archaeological Zone of Teotihuacan, Mexico. Material Issues in Art and Archaeology 852, 611e617. Van Strydonck, M., Dupas, M., Dauchot-Dehon, M., Pachiaudi, C., Marechal, J., 1986. The influence of contaminating (fossil) carbonate on the variations of d13C in mortar dating. Radiocarbon 28, 702e710. Van Strydonck, M., Dupas, M., Keppens, E., 1989. Isotopic fractionation of oxygen and carbon in lime mortar under natural environmental conditions. Radiocarbon 31, 610e618.